Modeled-Loss Triggers And Model Inputs

What a Modeled-Loss Trigger Measures

A catastrophe-linked contract can use a predefined catastrophe model to decide whether an event has crossed a contractual loss threshold. The model combines specified event characteristics with exposure information and financial terms to produce a modeled loss. The trigger is evaluated against that result—not necessarily against the cedent’s subsequently reported claims or a physical measurement such as wind speed.

The result depends on the inputs and model specification named in the contract. For example, a change in the event footprint, the exposure schedule, or the vulnerability assumptions can change the modeled loss even when the event being evaluated is otherwise described in the same way. A sound calculation therefore identifies the model run and its inputs before applying the trigger condition.

Definition: A modeled-loss trigger is a contractual condition evaluated using a loss value produced by a specified catastrophe model and a defined set of inputs. Let $Z$ denote the modeled trigger loss. The trigger may activate when $Z$ is greater than, or greater than or equal to, a stated threshold; the contract must specify which condition applies.

A simplified deterministic run can be described in four parts: the event characteristics supplied to the model, the exposed locations and their insured values, the modeled damage at each location, and the financial terms used to turn damage into the contract’s specified loss basis. The model’s output is not automatically the same as actual claims. It is an estimate generated for the contractual purpose and under the contractual rules.

Model-input principle: Before calculating $Z$, identify the event input, exposure schedule, damage assumptions, covered locations, financial terms, aggregation rule, and model version required by the contract. Do not substitute a different input or loss basis without contractual authority.

From Inputs to Location-Level Loss

For a simple location-level calculation, let $V_j$ be the insured value at location $j$, and let $r_j$ be the modeled damage ratio for the event and exposure characteristics at that location. The gross modeled damage is $V_jr_j$. If the contractual calculation applies a deductible $d_j$ and a per-location limit $u_j$, the location’s covered modeled loss is

L_j=\min\!\left(u_j,\max(0,V_jr_j-d_j)\right).

This formula assumes that the deductible is subtracted from modeled damage at each location, that negative results are floored at zero, and that the limit caps the resulting amount at that location. The contract may define different terms or a different order; use its stated definition. In particular, a model may report gross damage, insurer loss after policy terms, or another agreed loss measure. These are not interchangeable.

If the contract defines trigger loss as the sum of covered modeled losses across locations, then

Z=\sum_{j=1}^{n}L_j.

The aggregation rule matters. For example, applying separate location limits before summing can produce a different answer from applying one limit after aggregation. Likewise, deductibles applied separately at each location generally cannot be replaced by one portfolio-level deductible.

Worked Example: Recomputing Two Event Inputs

Method. Calculate the covered modeled loss at each location for each input set, sum the location results, and then compare each total with the stated trigger. A hypothetical contract covers two locations with insured values of $5.00 million and $8.00 million. The respective location deductibles are $0.20 million and $0.40 million. There are no additional limits. The contract pays a fixed $0.75 million if $Z\geq \$2.00$ million and pays zero otherwise. The supplied model outputs are damage ratios of 0.12 and 0.08 under input A, and 0.20 and 0.15 under input B.

Input A, location 1. Gross modeled damage is $5.00(0.12)=\$0.60$ million. After the deductible, the location loss is $0.60-0.20=\$0.40$ million.

Input A, location 2. Gross modeled damage is $8.00(0.08)=\$0.64$ million. After the deductible, the location loss is $0.64-0.40=\$0.24$ million. Therefore, $Z_A=0.40+0.24=\$0.64$ million. Since $0.64<2.00$, the trigger does not activate and the payment is $0.

Input B, location 1. Gross modeled damage is $5.00(0.20)=\$1.00$ million. After the deductible, the location loss is $1.00-0.20=\$0.80$ million.

Input B, location 2. Gross modeled damage is $8.00(0.15)=\$1.20$ million. After the deductible, the location loss is $1.20-0.40=\$0.80$ million. Thus, $Z_B=0.80+0.80=\$1.60$ million. Since $1.60<2.00$, the trigger also does not activate and the payment is $0.

Check and common mistake. Both location results are nonnegative, and each total is the sum of its two location losses. A tempting but incorrect approach is to use a different, unverified aggregate loss figure and compare that with the threshold. Recomputing from the supplied ratios gives $0.64 million for A and $1.60 million for B; neither reaches $2.00 million.

Why the Calculation Works

The location formula follows directly from the defined sequence of contractual adjustments. Begin with modeled damage $V_jr_j$. Subtract the deductible because the deductible is the portion of damage retained before the covered amount is determined. A floor at zero prevents a deductible from creating a negative covered loss. Finally, apply the limit by taking the smaller of the post-deductible amount and $u_j$. This gives

\text{covered loss} =\min(\text{limit},\max(0,\text{modeled damage}-\text{deductible})).

When the contract defines portfolio trigger loss as the sum of these location amounts, adding them follows from that definition: each $L_j$ is the location’s contribution to the covered modeled loss, so the portfolio total is $\sum_jL_j$. This aggregation is valid only for the stated basis. If the contract instead specifies a different measure—such as gross modeled damage or a differently aggregated loss—the calculation must follow that measure.

Fix the model run.
Record the contractual model version and event input. Confirm that the supplied event characteristics are the ones to be used.

Match inputs to exposures.
For each covered location, identify the insured value and the damage ratio produced under that run.

Apply the specified loss basis.
Apply deductibles, limits, and any other stated terms in the required order and at the required level.

Aggregate and test.
Calculate $Z$ using the contract’s aggregation rule, then compare it with the threshold using the stated strict or inclusive inequality.

Apply the stated payment rule.
Determine the payment from the contractual rule only after deciding whether the trigger condition is met.

Coverage Terms and Trigger Payments

The same event input can produce different trigger results for different exposure schedules or financial terms. Conversely, the same exposure schedule can produce different results when the model receives alternative event characteristics or damage assumptions. A comparison is meaningful only if it states what changed and holds other inputs fixed.

The contract also determines what happens after a trigger condition is satisfied. To keep the distinction clear, the examples here use explicit payment rules—such as a fixed amount if a threshold is reached. The modeled-loss calculation establishes the trigger result; the payment rule then converts that result into a contractual payment. Do not infer a payment amount from a trigger threshold alone.

Worked Example: Location Limits Change the Trigger Result

Method. Apply each deductible and limit separately before summing. Consider a hypothetical contract covering three locations. Values, damage ratios, deductibles, and per-location limits are all in millions of dollars: location 1 has $V_1=3.00$, $r_1=0.30$, $d_1=0.15$, and $u_1=0.50$; location 2 has $V_2=6.00$, $r_2=0.10$, $d_2=0.20$, and $u_2=0.35$; location 3 has $V_3=2.00$, $r_3=0.50$, $d_3=0.25$, and $u_3=0.80$. The contract pays $2.40 million if $Z\geq \$1.50$ million.

At location 1, gross modeled damage is $3.00(0.30)=\$0.90$ million. Subtract the deductible: $0.90-0.15=\$0.75$ million. Apply the limit: $L_1=\min(0.50,0.75)=\$0.50$ million.

At location 2, gross modeled damage is $6.00(0.10)=\$0.60$ million. After the deductible, $0.60-0.20=\$0.40$ million. The limit reduces this to $L_2=\min(0.35,0.40)=\$0.35$ million.

At location 3, gross modeled damage is $2.00(0.50)=\$1.00$ million. After the deductible, $1.00-0.25=\$0.75$ million. This is below the $0.80 million limit, so $L_3=\$0.75$ million.

The aggregate is $Z=0.50+0.35+0.75=\$1.60$ million. Because $1.60\geq1.50$, the trigger is met and the specified payment is $2.40 million.

Check and common mistake. The uncapped post-deductible losses sum to $0.75+0.40+0.75=\$1.90$ million. Using that as $Z$ would ignore the location limits. The correct total is $1.60 million, and it still exceeds the threshold. The payment is the stated fixed amount, not the amount by which $Z$ exceeds the threshold.

Worked Example: Alternative Model Versions and a Tiered Payment Rule

Method. Hold the exposure schedule and contract terms fixed, recalculate location losses under each set of model outputs, and apply the stipulated payment schedule. A hypothetical portfolio has two locations: location 1 has insured value $10.00 million, deductible $0.50 million, and limit $2.50 million; location 2 has insured value $4.00 million, deductible $0.20 million, and limit $2.00 million. Under model run A, the supplied damage ratios are 0.18 and 0.25. Under alternative run B, they are 0.28 and 0.40. The payment is $0 when $Z<\$3.00$ million, $1.00 million when $\$3.00\text{ million}\leq Z<\$4.00\text{ million}$, and $2.00 million when $Z\geq\$4.00$ million.

Run A. At location 1, gross damage is $10.00(0.18)=\$1.80$ million; subtracting $0.50 million gives $1.30 million, below the $2.50 million limit. Thus, $L_1=\$1.30$ million. At location 2, gross damage is $4.00(0.25)=\$1.00$ million; after the $0.20 million deductible, $L_2=\$0.80$ million, below its limit. Therefore, $Z_A=1.30+0.80=\$2.10$ million. Since $2.10<3.00$, the payment is $0.

Run B. At location 1, gross damage is $10.00(0.28)=\$2.80$ million; after the deductible, $2.80-0.50=\$2.30$ million, below the limit. At location 2, gross damage is $4.00(0.40)=\$1.60$ million; after the deductible, $1.60-0.20=\$1.40$ million, also below the limit. Thus, $Z_B=2.30+1.40=\$3.70$ million. This lies in the interval from $3.00 million inclusive to $4.00 million exclusive, so the payment is $1.00 million.

Check and common mistake. Both runs use the same exposures and financial terms; only the model outputs change. The alternative run raises $Z$ from $2.10 million to $3.70 million and moves the result into a different payment tier. Do not apply the upper tier merely because the threshold has been crossed: the specified intervals make the payment $1.00 million, not $2.00 million.

Model Governance and Common Exam Traps

Modeled-loss triggers make the model specification part of the contractual calculation. A result that cannot be reproduced from the stated event, exposures, assumptions, financial terms, and aggregation rule is difficult to verify and may be disputed. In practice, documentation should make clear which inputs are fixed by the contract, which data are supplied for the event, and how a change or correction to an input is handled.

The following errors can materially alter a trigger result:

Mixing loss bases. Gross modeled damage, loss after deductibles, and loss after both deductibles and limits are different quantities. Use the basis defined for $Z$.
Applying terms at the wrong level. Per-location deductibles and limits are applied to each location, not silently replaced by portfolio-level terms.
Changing multiple inputs without tracking them. When comparing runs, identify whether the difference comes from event characteristics, exposures, damage ratios, model version, or financial terms.
Misreading the boundary condition. “Greater than” excludes equality; “greater than or equal to” includes it. A result exactly at the threshold must be treated according to the written condition.
Confusing trigger activation with payment size. The trigger test identifies whether the contractual condition is met. The separate payment rule determines the amount.

Key takeaway: A modeled-loss trigger is reproducible only when the model run, exposure data, damage outputs, financial adjustments, aggregation rule, threshold inequality, and payment rule are all specified. Calculate location results on the stated basis first; aggregate and test only afterward.

Check Your Understanding

For each hypothetical contract, calculate the requested modeled loss or payment using only the stated inputs and terms.

A location has insured value $2.00 million, modeled damage ratio 0.25, deductible $0.10 million, and limit $0.40 million. A fixed $0.10 million is paid if $Z\geq\$0.35$ million. What are $Z$ and the payment? (Answer: $Z=\$0.40$ million; payment $0.10 million.)
Two locations have values $3.00 million and $5.00 million, ratios 0.20 and 0.10, deductibles $0.10 million and $0.20 million, and limits $0.50 million and $0.60 million. A fixed $0.30 million is paid if $Z\geq\$0.90$ million. Calculate $Z$ and the payment. (Answer: $Z=\$0.80$ million; payment $0.)
A two-location portfolio has values $4.00 million and $7.00 million, deductibles $0.10 million and $0.30 million, and no additional limits. Under input A, damage ratios are 0.15 and 0.10; under input B, they are 0.25 and 0.20. The contract pays $0.60 million if $Z\geq\$1.50$ million. Calculate both trigger losses and payments. (Answer: input A: $Z=\$0.90$ million, payment $0; input B: $Z=\$2.00$ million, payment $0.60 million.)
Three locations have, respectively, $(V_j,r_j,d_j,u_j)$ in millions of dollars: $(2.00,0.50,0.10,0.70)$, $(5.00,0.12,0.20,0.80)$, and $(3.00,0.20,0.10,0.50)$. A fixed $0.25 million is paid if $Z\geq\$1.50$ million. Calculate each location loss, $Z$, and payment. (Answer: $L_1=\$0.70$ million, $L_2=\$0.40$ million, $L_3=\$0.50$ million; $Z=\$1.60$ million; payment $0.25 million.)
A contract pays $0 when $Z<\$2.00$ million, $0.50 million when $\$2.00\text{ million}\leq Z<\$3.00$ million, and $1.20 million when $Z\geq\$3.00$ million. The calculated trigger loss is exactly $3.00 million. What payment applies, and why? (Answer: $1.20 million, because the upper tier includes equality at $3.00 million.)

What You'll Learn